Measurements of the canine skull are of main interest in brachycephalic dogs. They are suspected to have diseases associated with the shortened nose, which are summarized under the term brachycephalic syndrome. The relative length of the nose has become an important feature in breeding and research, although a defi nitive correlation to the brachycephalic syndrome still has not been proven (Knecht, 1979; Harvey, 1982; Aron und Crowe, 1985; Koch et al., 2003). Classifi cations of skull types can be made by two methods. The traditional phenotypic assignment into brachycephalic, mesaticephalic and dolichocephalic is based on the assumption, that dogs of a specifi c breed have more or less uniform skull shapes. However, alterations due to breeding and differences be tween individual dogs of the same breed must lead to modifi cations. Therefore and for a more subtle classifi cation, skull measurements were introduced (Brehm et al., 1985; Evans, 1993; Regodon et al., 1993), The existing indexation methods from Evans (1993) and Brehm et al. (1985) are based on cadaver studies. They classify brachycephalic dogs as having a cranial length to facial length ratio (LL-index) of > 1.60 or a skull length to skull width ratio (LW-index) of < 1.44. However, the authors also mentioned the considerable range of those indices within a breed. Regodon et al. (1993) were the fi rst to propose a classifi cation based on radiographs of canine skulls with the craniofacial angle (CFA) as a criterion. The CFA is defi ned by a line along the base of the skull, and another line along the hard palate. The CFA of 10 dogs from 5 breeds were retrieved and it was found that the ranges clearly overlapped. The existing indices have their limitations for clinical use, either being based on cadavers (LL-index, LW-index) or not resulting in meaningful assignments of the individual skull measurement to the phenotypic appearance of the respective breed (CFA). We therefore propose a possible new skull index (S-index), which can be easily derived from radiographs of living animals and should refl ect the phenotype of each individual. Such an index has also the potential to be linked to airstream parameters (Wiestner et al., 2007) and to the clinical manifestation of respiratory problems of each individual. It offers an insight to different skull shapes between dogs of the same breed. We also propose a boundary value, which should have the potential to classify between brachycephalic and non-brachycephalic skull shapes. The new S-index is an approximation to the LL-index, retrieved form cadavers by Brehm et al. (1985). The index however was taken from radiographs in dorsoventral direction (Fig. 1). It is calculated by the length of the facial skull divided by the length of the cerebrum. In this context the length of the facial skull was defi ned as the distance from the rostral border of the incisal bone to the rostral border of the cavum cranii. The length of the cerebrum was measured from the rostral border of the cavum cranii to the caudal outline of the occipital bone, beginning at the dorsal border of the great foramen. The S-index was compared to the LW index and the CFA on 57 dogs of different breed, sex, age and size. They were presented to the University of Zurich for planned surgical procedures (mainly neutering). The owners agreed to the procedures necessary for this study. From all dogs, radiographs of the skull in dorsoventral (hard palate parallel to the table, Fig. 1) and laterolateral direction (hard palate perpendicular to the table, Fig. 2) were taken. The Sand LW-index (dorsoventral views) and CFA (laterolateral views) were determined (Tab. 1). A boundary value (BV) for the S-index was evaluated which allows for a separation of the brachycephalic and the non-brachycephalic breeds by the following method. First, the dogs in the present study were classifi ed on the basis of the established separating BV for the LW-index (Brehm et al., 1985). Secondly, the BV of S-index was determined such that the highest score in coincidence of classifi cation occurred for both, the brachycephalic and non-brachycephalic breeds . In this manner, it was found that all dogs having an S-index equal or below the BV of 1.25 are considered to have a brachycephalic skull shape (Tab. 1). In parallel the respective CFA were listed.